DBD::Sys - System tables interface via DBI


  use DBI;
  my $dbh = DBI->connect('DBI::Sys:');
  my $st  = $dbh->prepare('select distinct * from filesystems join filesysdf on mountpoint');
  my $num = $st->execute();
  if( $num > 0 )
      while( my $row = $st->fetchrow_hashref() )
          # ...


DBD::Sys is a so called database driver for DBI designed to request information from system tables using SQL. It's based on SQL::Statement as SQL engine and allows to be extended by DBD::Sys::Plugins.


Of course, a DBD requires DBI to run. Further, SQL::Statement as SQL engine is required, Module::Pluggable to manage the plugin's and Module::Build for installation. Finally, to speed up some checks, Params::Util is needed.

All these modules are mandatory and DBD::Sys will fail when they are not available.

To request system information, existing modules from CPAN are used - there are available ones to provide access to some system tables. These modules are optional, but recommended. It wouldn't make much sense to use DBD::Sys without the ability to access the tables from the (operating) system.

To get an overview which dependencies are there, please check the plugins or take a look into META.yml.



We chose Module::Build installation, because not every system has a suitable make utility - but at least everyone who's using perl modules has a running perl. So installing can be done after extracting

  gzip -dc DBD-Sys-${VERSION}.tar.gz | tar xvf -

without too much extra effort:

  1  cd DBD-Sys-${VERSION}
  2  perl Build.PL
  3  ./Build
  4  ./Build test
  5  ./Build install

If you want to skip the tests (not recommended), you can skip over lines 3 and 4.

Fetching data

To retrieve data, you can use the following example:

    my $dbh = DBI->connect('DBI:Sys:');
    $st  = $dbh->prepare( 'SELECT DISTINCT username, uid FROM pwent WHERE username=?' );
    $num = $st->execute(getlogin() || $ENV{USER} || $ENV{USERNAME});
    while( $row = $st->fetchrow_hashref() )
        printf( "Found result row: uid = %d, username = %s\n", $row->{uid}, $row->{username} );

Error handling

Errors while processing statements are handled via DBI - read DBI documentation, especially the err and errstr documentation, if you're not familiar with error handling in DBI.

Errors while modifying attributes, calling driver methods etc. are reported by throwing an exception using Carp.


Each table implementor can request configurable meta data attributes. They will be accessible via the database handle:

    print $dbh->{"sys_" . lc $table . "_" . $attr}, "\n";
    # e.g.
    print DBI->neat( $dbh->{sys_filesysdf_blocksize} ), "\n";


This module does not support any changes to the provided tables in order to prevent inconsistent data.

The design of the plugins makes it less predictable what columns are provided in the end. Well, at least those columns from the tables provided by the DBD::Sys::Plugin::Meta and DBD::Sys::Plugin::Any will be available, even if they are not filled with data when the appropriate module is missing (e.g. if Sys::Filesystem is not available, the table filesystems gets the columns provided by DBD::Sys::Plugin::Any::Filesys, but no data at all).

All additional table implementors must use the same primary key as all other implementors. To stay at the example of filesystems, the primary key is mountpoint - and if any additional module provides another implementation (with data from another module than Sys::Filesystem), it needs to ensure that the column mountpoint is provided and unique. Additionally it must return mountpoint as primary key when it's method get_primary_key is invoked.


    Jens Rehsack                        Alexander Breibach
    CPAN ID: REHSACK          


This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

The full text of the license can be found in the LICENSE file included with this module.


Free support can be requested via regular CPAN bug-tracking system at There is no guaranteed reaction time or solution time, but it's always tried to give accept or reject a reported ticket within a week. It depends on business load. That doesn't mean that ticket via rt aren't handles as soon as possible, that means that soon depends on how much I have to do.

Business and commercial support should be acquired from the authors via preferred freelancer agencies.


perl(1), DBI, Module::Build, Module::Pluggable, Params::Util, SQL::Statement.